Thushan Ganegedara
Machine Learning Engineer at Towards Data Science
Sydney, New South Wales, Australia
Summary
👤
Senior🎓
Top SchoolThushan Ganegedara is a Machine Learning Engineer with 14 years of experience, currently working on Trust & Safety ML at TikTok and formerly a senior ML engineer at Canva. He holds a PhD in AI from the University of Sydney, has authored two TensorFlow books, and teaches popular courses and workshops—reaching thousands of learners via DataCamp and Medium. His hands-on open-source work includes Keras attention layers and practical CV/NLP notebooks, showing a focus on model architecture, training loops and deployable tooling. A frequent conference speaker and technical writer, he bridges research and production, especially in domain-specific NLP and scalable ML infrastructure. Less obvious: he combines academic research on adaptive model design with pragmatic implementations used in tutorials and repos that emphasize reproducibility and environment setup.
14 years of coding experience
11 years of employment as a software developer
BSc (Hons) Computer Sceince, BSc (Hons) Computer Sceince at University of MoratuwaThe University of SydneyNatural Language Processing Nanodegree Information Technology, Natural Language Processing Nanodegree Information Technology at UdacityMathematics, Mathematics at Kingswood College, Sri Lanka
English, Sinhalese
Stackoverflow
11,155reputation
853kreached
344answers
1question
signal-processing
top-5%
neural-network
top-5%
tensorflow
top-1%
deep-learning
top-1%
machine-learning
top-1%
keras
top-1%
Github Skills (27)
python10
attention-mechanism10
machine-learning10
rnn-model10
n10
mask-rcnn10
keras10
lstm10
deep-learning10
tensorflow10
computer-vision10
faster-rcnn10
signal-processing9
word2vec9
importerror9
Programming languages (6)
Github contributions (5)
thushv89/attention_keras
Mar 2019 - Jul 2022
Keras Layer implementation of Attention for Sequential models
Role in this project:
ML Engineer
ML EngineerContributions:1 release, 53 commits, 18 PRs in 3 years 4 months
Contributions summary:Thushan contributed to the implementation of an attention layer for a Keras-based sequential model, as evidenced by the commit messages and code diffs. They fixed bugs, updated documentation, and refactored the code. Further contributions included the creation of a utility module and the addition of both bidirectional and variable sequence length support to the model. The user also dockerized the project and updated the environment setup.
rnndeep-learninglstmtensorflowsequential-models
thushv89/exercises_thushv_dot_com
Oct 2017 - Jul 2022
Here lies all the exercises I implement and share in my website
Role in this project:ML Engineer
ML EngineerContributions:34 commits, 32 pushes, 1 branch in 4 years 9 months
Contributions summary:Thushan's primary contribution revolves around implementing machine learning models for computer vision and natural language processing tasks. They worked on developing a Convolutional Neural Network (CNN) model for MNIST, a style transfer network using VGG-16 architecture, and a word2vec model. The commits indicate a focus on the model architecture, training loops, and data preprocessing steps. The user added a Bayesian Linear Regression notebook and implemented a decision tree.
nlpdeep-learningmachine-learningcomputervisionshare
Find and Hire Top DevelopersWe’ve analyzed the programming source code of over 60 million software developers on GitHub and scored them by 50,000 skills. Sign-up on Prog,AI to search for software developers.